In this paper, we consider an improved and efficient algorithm for the multiuser detection (MUD) in direct sequence/code division multiple access (DS/CDMA) communication systems. The optimum detector for MUD is the maximum likelihood (ML) detector, but its complexity is very high and it calls for an optimization problem that involves an exhaustive search. Consequently, there has been considerable interest in suboptimal multiuser detectors with less complexity and reasonable performance. The idea of reducing the complexity by sub-optimum detectors was efficient and applicable but had some defects. The idea proposed in this paper is the restriction of the search space. This idea comprises using a sub-optimum detector such as ordinary genetic algorithm for which the size of the search space is confined to a predetermined region that is smaller than the whole search space. This limited region includes the objective bit sequence. Then by using depth-first tree search algorithm, we could find the optimum bit stream. We compare our algorithm to the ML detector, the genetic algorithm with conventional detector, and the ant colony optimization (ACO) detector which have been used for MUD in DS/CDMA. Simulation results show that the performance of this algorithm is near the optimal detector with very low complexity, and works better in comparison to other algorithms.